Joint binning label generation method and device for risk portrait construction

Abstract

The invention provides a joint binning label generation method and device for risk portrait construction, and the method comprises the steps: S1-S3, dividing all features into a plurality of sub-feature sets, selecting a plurality of features from all the sub-feature sets according to the importance values of the features, and constructing an importance representative feature set; s4, performing binning calculation on the importance representative feature set to obtain a binning value of each feature corresponding to each target in the importance representative feature set, and constructing a binning feature set according to all the binning values; s5, combining the features in the binning feature set to obtain a combined binning feature set; and step S6, calculating the woe value of each feature corresponding to each target in the joint binning feature set. In a word, the method can generate a plurality of tags for improving the accuracy of the risk portrait and the corresponding tag value of each target.

Description

Technical Field

[0001] The present invention belongs to the field of risk profiling, and specifically relates to a method and device for generating joint binning labels for risk profiling construction. Background Art

[0002] Risk profiling is used in multiple fields, and the key to the accuracy of risk profiling lies in whether the risk label can reflect the risk status of the subject in a certain dimension.

[0003] At present, the subject risk profiling mostly adopts a rule-driven method based on expert experience, integrating various features into risk labels in order to assess the subject risk. This method is highly interpretable and logically clear, but it also has significant limitations, as shown below:

[0004] 1) Dimensionality curse caused by high-dimensional features: The data to be analyzed is usually high-dimensional and diverse, and the relationship between features is complex and non-explicit. Rule-based methods are difficult to effectively extract key information in high-dimensional feature space, and often result in low-quality label generation due to limited rule coverage or overly simple design. In addition, the computational cost brought by high-dimensional features also limits the widespread application of rule-based methods.

[0005] 2) Lack of systematic label generation process: Rule-driven methods usually do not have a unified feature selection, feature expansion, and risk label generation framework, and rely on customized designs for different scenarios, which lacks systematicity. This lack of standardized generation process increases the difficulty of implementation and maintenance, while reducing the interpretability of the model and the reliability of risk assessment.

[0006] 3) Insufficient feature selection and importance assessment: The rule-driven approach relies on the experience and knowledge of domain experts to define feature selection rules, feature combination methods, and risk assessment indicators. This approach may ignore the interactions and potential associations between different features, resulting in the omission of key information, thereby reducing the predictive performance of the model.

[0007] 4) Single binning processing method and lack of interactive analysis capabilities: Current binning technologies mostly use independent single-feature binning, which is difficult to reflect the interactive relationship between multiple features. Single-feature binning may lead to insufficient multi-dimensional understanding of risk groups and affect the accurate characterization of complex risk states. Summary of the invention

[0008] The present invention is made to solve the above-mentioned problems, and its purpose is to provide a method and device for generating joint bin labels for risk profiling.

[0009] The present invention provides a joint binning label generation method for risk profile construction, which is used to generate label results related to risk profiles corresponding to each target based on risk-related data, wherein the risk-related data includes values ​​of multiple different features corresponding to each target, and risk judgment labels corresponding to each target, and has such characteristics, including the following steps: step S1, dividing all features into multiple categories, generating sub-feature sets corresponding to each category, and the sub-feature sets including at least one feature and all values ​​corresponding to the feature; step S2, for each sub-feature set, calculating the importance value of each feature according to the number of features in the sub-feature set and the values ​​corresponding to the features; step S3, selecting multiple features from all sub-feature sets according to the importance values, and constructing Importance representative feature set; step S4, binning calculation on the importance representative feature set to obtain the binning value of each feature corresponding to each target in the importance representative feature set, and construct a binning feature set based on all the binning values; step S5, combine each feature in the binning feature set to obtain a combined binning feature set; step S6, binning calculation on the combined binning feature set to obtain the binning value of each feature corresponding to each target in the combined binning feature set, and calculate the woe value of each feature corresponding to each target in the combined binning feature set based on all the binning values ​​and risk judgment labels corresponding to the combined binning feature set, wherein the label result of the target is each feature of the target in the combined binning feature set and its corresponding woe value.

[0010] The combined binning label generation method for risk profiling provided by the present invention may also have the following features: wherein, in step S4, the calculation expression of the binning feature set is: In the formula is the binned feature set, q is the total number of features in the feature set representing importance, is the bin value in the bin value of feature i corresponding to target k, and n is the total number of targets.

[0011] The method for generating joint binning labels for risk profiling provided by the present invention may also have the following features: wherein, in step S5, the calculation expression of the joint binning feature set is: UnionBins(x)=[χ1||χ2,χ1||χ3,…,χ n-1 ||χ n ],x=[χ1,χ2,…,χ n ], where X unionbin is the joint binning feature set, χ n is the nth feature in the feature vector set x, and || is the set concatenation symbol.

[0012] In the joint binning label generation method for risk profile construction provided by the present invention, it can also have the following characteristics: wherein, binning the feature set to obtain the binning value includes the following steps: step T1, generating multiple candidate splitting thresholds for the feature set, and calculating the splitting value corresponding to each candidate splitting threshold; step T2, taking the candidate splitting threshold corresponding to the largest splitting value as the optimal splitting point; step T3, repeating steps T1 to T2 until the number of optimal splitting points reaches a preset value; step T4, calculating the binning boundaries based on all the optimal splitting points; step T5, according to the binning boundaries, calculating the bins value and bin0 value of each feature corresponding to each target in the feature set as the binning value.

[0013] In the joint binning label generation method for risk profile construction provided by the present invention, there may also be such a feature: wherein the calculation expression of the split value is: H(D j )=-∑ c∈C p(c)log2p(c), In the formula, IG(D j ,θ) is the feature set D j The corresponding split value of the candidate split threshold θ, |D| is the number of samples in D, is the feature set D j The value of feature i corresponding to target k, v k is the target k, y k is the risk judgment label corresponding to target k, C is the feature set D j The total number of features in the feature set D, p(c) is the number of features c in the feature set D j The probability of the bin boundary is calculated as: In the formula for The set of bin boundary values, To sort all optimal split points in ascending order, the calculation expressions for bins value and bin0 value are: In the formula is the feature set D j The value of feature i corresponding to target k in , for The left boundary value of the u-th binning interval.

[0014] In the joint binning label generation method for risk profile construction provided by the present invention, it can also have the following characteristics: wherein, in step S2, according to the information gain of the feature on the node of the tree of the LightGBM model, the corresponding importance value is calculated, and the calculation expression of the importance value is: In the formula is the importance value of feature i in sub-feature set j, p is the total number of features in sub-feature set j, M is the total number of trees in the LightGBM model, N m is the set of nodes of the mth tree, is the indicator function, ΔL m,n is the information gain of feature i on node n of the mth tree. When node n is split using feature i, When node n does not use feature i for splitting,

[0015] The joint binning label generation method for risk profile construction provided by the present invention may also have the following characteristics: wherein, in step S3, the calculation expression of the importance representative feature set is: Where X lgbm_represent is the representative feature set of importance, m is the total number of sub-feature sets, X j,lgbm_top In order to filter the importance feature set corresponding to the j-th sub-feature set, all features of the sub-feature set j are sorted from large to small according to the importance value, and a preset number of features are selected from the beginning of the sequence to construct an importance filtering feature set.

[0016] In the joint binning label generation method for risk profile construction provided by the present invention, it can also have the following characteristics: wherein, in step S6, according to the bin0 value in the binning value of the feature corresponding to the target, the values ​​of the features corresponding to all targets are divided into multiple groups, the values ​​of the features corresponding to all targets in the group have the same bin0 value, the risk judgment label is a binary classification label, and the calculation expression of the woe value is: In the formula is the woe value of feature i corresponding to target k in the joint bin feature set, N bad N is the total number of risk judgment labels with bad labels corresponding to all values ​​in the group where the value of feature i corresponding to target k in the joint binning feature set belongs, good For target k in sub-feature set D j The total number of risk judgment labels corresponding to all values ​​in the group where the corresponding feature i value belongs to the good category, N total is the total number of values ​​of feature i corresponding to target k in the joint bin feature set.

[0017] The present invention also provides a joint binning label generation device for risk portrait construction, which is used to generate label results related to risk portraits corresponding to each target based on risk-related data. The risk-related data includes the values ​​of multiple different features corresponding to each target, and the risk judgment label corresponding to each target, which has the following characteristics, including: a subset division module, which divides all features into multiple categories, and generates sub-feature sets corresponding to each category, and the sub-feature sets include at least one feature and all values ​​corresponding to the feature; an importance value calculation module, for each sub-feature set, according to the number of features in the sub-feature set and the values ​​corresponding to the features, calculates the importance value of each feature; an importance representative feature set generation module, which selects multiple features from all sub-feature sets according to the importance values, and constructs an important representative feature set. The importance representative feature set; the binning feature set generation module, performs binning calculation on the importance representative feature set, obtains the binning value of each feature corresponding to each target in the importance representative feature set, and constructs the binning feature set according to all the binning values; the joint binning feature set generation module, combines each feature in the binning feature set to obtain the joint binning feature set; the woe calculation module, performs binning calculation on the joint binning feature set, obtains the binning value of each feature corresponding to each target in the joint binning feature set, and calculates the woe value of each feature corresponding to each target in the joint binning feature set according to all the binning values ​​and risk judgment labels corresponding to the joint binning feature set, wherein the label result of the target is each feature of the target in the joint binning feature set and its corresponding woe value.

[0018] Functions and Effects of the Invention

[0019] According to the method and device for generating joint bin labels for risk profile construction involved in the present invention, by dividing the features into sub-feature sets, and calculating and filtering from each sub-feature set according to the corresponding importance values ​​to obtain multiple features to construct an importance representative feature set, and then binning the importance representative feature set and combining to obtain a joint bin feature set, and finally calculating the woe value of each target for the features in the joint bin feature set, thereby obtaining a label result. Therefore, the method and device for generating joint bin labels for risk profile construction of the present invention can generate multiple labels and corresponding label values ​​for each target that improve the accuracy of risk profiles. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a block diagram of a joint binning label generation device for risk profile construction in an embodiment of the present invention;

[0021] Figure 2 is a schematic diagram of a process for calculating bin values ​​in an embodiment of the present invention;

[0022] Figure 3 It is a flowchart of a joint binning label generation method for risk profile construction in an embodiment of the present invention;

[0023] Figure 4 is a radar schematic diagram of a high-risk user portrait in an embodiment of the present invention;

[0024] Figure 5 is a radar schematic diagram of a low-resolution user portrait in an embodiment of the present invention;

[0025] Figure 6 Schematic diagram of visualizing a scorecard in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following embodiments, in conjunction with the accompanying drawings, specifically illustrate the method and device for generating joint bin labels for risk profiling of the present invention.

[0027] This embodiment provides a joint binning label generation device for risk profile construction, which is used to generate label results related to risk profiles for each target according to risk-related data. The risk-related data includes the values ​​of multiple different features corresponding to each target, and the risk judgment label corresponding to each target. The risk judgment label is a binary classification label.

[0028] In this embodiment, the joint bin label generation device for risk profiling is applied to the financial field, and the risk-related data is financial transaction data D = [VXY]. Where V = {v1, ..., v n} represents a set of trading accounts, each trading account is a target, for each v i ∈V has a binary label whether it is illegal, that is, the line judgment label y i ∈Y, n is the number of trading accounts, and the feature space constructed from the perspectives of statistical calculation, complex rules, expert knowledge, and intelligent algorithms is X, where X = {X1, X2, …, X d}, X i ={ i,k |k=1,2,…,n}, represents the feature χ i The feature set of i,k Indicates the feature χ corresponding to the kth trading account i In other embodiments, the combined bin label generation device for risk profile construction can be applied to other fields, such as food safety, where the target is a specified food, and the risk-related data is production data related to the specified food, such as the manufacturer.

[0029] Figure 1It is a block diagram of a joint binning label generation device for risk profiling construction in an embodiment of the present invention.

[0030] like Figure 1 As shown, the joint binning label generation device 100 for risk profiling construction includes a subset partitioning module 11, an importance value calculation module 12, an importance representative feature set generation module 13, a binning feature set generation module 14, a joint binning feature set generation module 15, a woe calculation module 16 and a control module 17 for controlling the operation of the above modules.

[0031] The subset division module 11 is used to divide all features into multiple categories and generate a sub-feature set corresponding to each category. The sub-feature set includes at least one feature and all values ​​corresponding to the feature.

[0032] In this embodiment, the subset division module 11 divides the information into m categories, that is, obtains the sub-feature set X 1 ,X 2 ,…,X m , and satisfies

[0033] The importance value calculation module 12 is used to calculate the importance value of each feature in each sub-feature set according to the number of features in the sub-feature set and the values ​​corresponding to the features.

[0034] Among them, the importance value calculation module 12 includes a LightGBM model, and the corresponding importance value is calculated according to the information gain of the feature on the node of the tree of the LightGBM model. The calculation expression of the importance value is:

[0035]

[0036] In the formula is the importance value of feature i in sub-feature set j, p is the total number of features in sub-feature set j, M is the total number of trees in the LightGBM model, N m is the set of nodes of the mth tree, is the indicator function, ΔL m,n is the loss reduction when feature i is used as the splitting loss at node n of the mth tree, i.e., information gain. When node n is split using feature i, When node n does not use feature i for splitting,

[0037] The loss function corresponding to the LightGBM model in this embodiment is the binary cross entropy loss L, and its calculation expression is:

[0038]

[0039] Where N is the total number of samples, y i is the risk judgment label of the i-th sample, is the predicted probability of the LightGBM model for the i-th sample.

[0040] The importance representative feature set generation module 13 is used to select multiple features from all sub-feature sets according to importance values ​​to construct an importance representative feature set.

[0041] Among them, the calculation expression of the importance representative feature set is:

[0042]

[0043] Where X lgbm_reporesent is the representative feature set of importance, m is the total number of sub-feature sets, X j,lgbm_top Filter the feature set for the importance corresponding to the j-th sub-feature set.

[0044] All features of sub-feature set j are sorted from large to small according to the importance value, and a preset number of features are selected from the beginning of the sequence to construct an importance screening feature set. The calculation expression of the importance screening feature set is:

[0045] X j,lgbm_top =argsort(|lgbm_importance j |)[:1],

[0046]

[0047] In the formula, argsort(|lgbm_importance j |) is to arrange the importance values ​​of the sub-feature set j in descending order according to the absolute value, that is, in this embodiment, the order is based on the absolute value of the importance value, and [:1] is a preset number.

[0048] The binning feature set generation module 14 is used to perform binning calculations on the importance representative feature set, obtain binning values ​​of each feature corresponding to each target in the importance representative feature set, and construct a binning feature set based on all binning values.

[0049] Figure 2 It is a schematic diagram of the process of calculating the bin value in an embodiment of the present invention.

[0050] like Figure 2 As shown, binning the feature set to obtain the bin value includes the following steps:

[0051] Step T1: generate multiple candidate splitting thresholds for the feature set, and calculate the splitting value corresponding to each candidate splitting threshold.

[0052] The calculation expression of the split value is:

[0053]

[0054] H(D j )=-∑ c∈C p(c)log2p(c),

[0055]

[0056] In the formula, IG(D j ,θ) is the sub-feature set D j The corresponding split value of the candidate split threshold θ, |D| is the number of samples in D, and a sample includes the values ​​of all features corresponding to a target in the set D, is the sub-feature set D j The value of feature i corresponding to target k, v k is the target k, y k is the risk judgment label corresponding to target k, C is the sub-feature set D j The total number of features in D, p(c) is the number of features c in the sub-feature set D j probability.

[0057] Step T2: taking the candidate splitting threshold corresponding to the maximum splitting value as the optimal splitting point.

[0058] Step T3, repeating steps T1 to T2 until the number of optimal splitting points reaches a preset value. In this embodiment, iterating two layers generates at most four optimal splitting points.

[0059] Step T4, calculating the bin boundaries based on all the optimal splitting points.

[0060] The calculation expression of the bin boundary is:

[0061]

[0062] In the formula for The set of bin boundary values, To sort all the optimal splitting points in ascending order. In this embodiment, negative infinity -∞ and positive infinity ∞ are used to ensure that the final binning boundary allows all values ​​to be covered.

[0063] Step T5, according to the binning boundaries, calculate the bins value and bin0 value of each feature corresponding to each target in the feature set as the binning value.

[0064] Among them, the calculation expressions of bins value and bin0 value are:

[0065]

[0066] In the formula is the sub-feature set D j The value of feature i corresponding to target k in , for The left boundary value of the u-th binning interval.

[0067] Among them, the calculation expression of the bin feature set is:

[0068]

[0069] In the formula is the binned feature set, q is the total number of features in the feature set representing importance, is the bin value in the bin value of feature i corresponding to target k, and n is the total number of targets.

[0070] In this embodiment, the binning feature set generation module 14 uses the importance representative feature set as the feature set and calculates the corresponding binning value.

[0071] The combined binning feature set generation module 15 is used to combine the features in the binning feature set to obtain a combined binning feature set.

[0072] Among them, in step S5, the calculation expression of the joint bin feature set is:

[0073]

[0074] UnionBins(x)=[χ1||χ2,χ1||χ3,…,χ n-1 ||χ n ],

[0075] x=[χ1,χ2,…,χ n ],

[0076] Where X unionbin is the joint binning feature set, χ n is the nth feature in the feature vector set x, and || is the set concatenation symbol.

[0077] The woe calculation module 16 is used to perform bin calculation on the joint bin feature set, obtain the bin value of each feature corresponding to each target in the joint bin feature set, and calculate the woe value of each feature corresponding to each target in the joint bin feature set according to all the bin values ​​and risk judgment labels corresponding to the joint bin feature set. Among them, the label result of the target is each feature of the target in the joint bin feature set and its corresponding woe value.

[0078] In this embodiment, the woe calculation module 16 uses the joint binning feature set as the feature set and calculates the corresponding binning value.

[0079] Among them, the woe calculation module 16 divides the values ​​of the features corresponding to all targets into multiple groups according to the bin0 value of the feature corresponding to the target The values ​​of the features corresponding to all targets in the group have the same bin0 value, that is, if The target corresponding feature

[0080] The calculation expression of woe value is:

[0081]

[0082] In the formula For target k in sub-feature set D j The woe value of feature i corresponding to N bad For target k in sub-feature set D j The total number of risk judgment labels whose labels are bad categories in the group where the corresponding feature i value belongs, N good For target k in sub-feature set D j The total number of risk judgment labels corresponding to all values ​​in the group where the corresponding feature i value belongs to the good category, N total For target k in sub-feature set D j The total number of values ​​of the corresponding feature i in . In this embodiment, the risk judgment label is 0 or 1, 0 represents a good category, and 1 represents a bad category. In other embodiments, other forms can be used to replace 0 and 1 to distinguish between good and bad categories.

[0083] The control module 17 stores a control program for controlling the operation of each module.

[0084] The following describes the process of using the joint bin label generation device 100 for risk profile construction to perform a joint bin label generation method for risk profile construction in conjunction with the accompanying drawings.

[0085] Figure 3 It is a flowchart of a joint binning label generation method for risk profiling construction in an embodiment of the present invention.

[0086] like Figure 3 As shown in FIG. 1 , the joint binning label generation method for risk profile construction includes the following steps:

[0087] Step S1: using the subset division module 11 to divide all features into multiple categories, generating a sub-feature set corresponding to each category, wherein the sub-feature set includes at least one feature and all values ​​corresponding to the feature.

[0088] Step S2: The importance value calculation module 12 calculates the importance value of each feature in each sub-feature set according to the number of features in the sub-feature set and the value corresponding to the feature.

[0089] Step S3: The importance representative feature set generation module 13 is used to select multiple features from all sub-feature sets according to the importance values ​​to construct an importance representative feature set.

[0090] Step S4, using the binning feature set generation module 14 to perform binning calculations on the importance representative feature set, obtain the binning values ​​of each feature corresponding to each target in the importance representative feature set, and construct a binning feature set based on all the binning values.

[0091] Step S5: Use the combined binning feature set generation module 15 to combine the features in the binning feature set to obtain a combined binning feature set.

[0092] Step S6, use the Woe calculation module 16 to perform bin calculation on the joint bin feature set to obtain the bin value of each feature corresponding to each target in the joint bin feature set, and calculate the Woe value of each feature corresponding to each target in the joint bin feature set based on all the bin values ​​and risk judgment labels corresponding to the joint bin feature set.

[0093] In this embodiment, a radar chart of the user portrait of the target can be constructed based on the label results of the target.

[0094] Figure 4 It is a radar schematic diagram of a high-risk user portrait in an embodiment of the present invention.

[0095] Figure 5 It is a radar schematic diagram of a low-risk user portrait in an embodiment of the present invention.

[0096] like Figure 4 and Figure 5 As shown in the figure, the label result includes 8 features, namely union bin labels and their corresponding values, namely union_bin_1, union_bin_2, union_bin_3, union_bin_4, union_bin_5, union_bin_6, union_bin_7 and union_bin_8. It can be seen that the area occupied by abnormal subjects, namely high-risk users, is large, while the area occupied by normal subjects, namely low-risk users, is small, which can intuitively show that the multi-dimensional risk level of the subject risk can be accurately characterized by the joint bin labels.

[0097] Based on the existing scoring model, this embodiment extracts the characteristic coefficient and intercept, sets the benchmark score and the score PDO for each doubling of the odds, calculates the factor weight and offset, and creates a score card containing the characteristics, coefficients, sample scores and predicted probabilities. The scores are divided into several equidistant scoring intervals, and the number and proportion of positive and negative samples in each scoring interval are calculated to intuitively reflect the distribution of positive and negative samples in each scoring interval.

[0098] Figure 6 Schematic diagram of visualizing a scorecard in an embodiment of the present invention.

[0099] like Figure 6 As shown, the horizontal axis is the sample score interval value score_bins, the vertical axis of the histogram is the number of samples in the score interval Count, and the vertical axis of the line graph is the bad sample ratio Bad Rate in the score interval. It can be seen that the score card based on the model label, i.e., the target result of this embodiment, intuitively shows the distinction between positive and abnormal by each module, and realizes a strong association between abnormal subjects and risk scores.

[0100] Functions and Effects of the Embodiments

[0101] According to the joint binning label generation method and device for risk profiling involved in this embodiment, by dividing the features into sub-feature sets, and calculating and filtering out multiple features from each sub-feature set according to the corresponding importance values ​​to construct an importance representative feature set, the importance representative feature set is then binned and combined to obtain a joint binning feature set, and finally the woe value of each target is calculated for the features in the joint binning feature set, thereby obtaining a label result. In summary, on the one hand, this method reduces irrelevant or redundant features, reduces feature complexity, and alleviates the dimensionality disaster problem by constructing a representative feature set of importance. On the other hand, through binning statistics and feature cross-combination, a label result that can reflect the interactive relationship of multi-dimensional features is generated, the feature interaction analysis capability is enhanced, and the model's learning ability for complex data patterns is improved.

[0102] Those skilled in the art should understand that the present invention is not limited to the above embodiments, and the above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, and these changes and improvements fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A joint binning label generation method for risk profile construction, which is used to generate label results related to risk profiles corresponding to each target according to risk-related data, wherein the risk-related data includes values ​​of multiple different features corresponding to each target and risk judgment labels corresponding to each target, characterized in that: The following steps are involved: Step S1, dividing all the features into multiple categories, generating a sub-feature set corresponding to each category, wherein the sub-feature set includes at least one feature and all the values ​​corresponding to the feature; Step S2, for each of the sub-feature sets, according to the number of the features in the sub-feature set and the values ​​corresponding to the features, calculate the importance value of each of the features; Step S3, selecting a plurality of the features from all the sub-feature sets according to the importance values, and constructing an importance representative feature set; Step S4, performing bin calculation on the importance representative feature set to obtain bin values ​​of each feature corresponding to each target in the importance representative feature set, and constructing a bin feature set based on all the bin values; Step S5, combining the features in the binned feature set to obtain a combined binned feature set; Step S6, performing the binning calculation on the joint binning feature set to obtain the binning values ​​of each feature corresponding to each target in the joint binning feature set, and calculating the woe value of each feature corresponding to each target in the joint binning feature set based on all the binning values ​​corresponding to the joint binning feature set and the risk judgment label. Among them, the label result of the target is each of the features of the target in the joint bin feature set and its corresponding woe value.

2. The method for generating joint bin labels for risk profile construction according to claim 1, characterized in that: in, In step S4, the calculation expression of the binning feature set is: In the formula is the binned feature set, q is the total number of features in the feature set representing importance, is the bin value in the bin value of feature i corresponding to target k, and n is the total number of targets.

3. The joint binning label generation method for risk profile construction according to claim 1 is characterized in that: in, In step S5, the calculation expression of the joint binning feature set is: UnionBins(x)=[χ1||χ2,χ1||χ3,…,x n-1 ||x n ], x=[χ1,x2,…,x n ], Where X unionbin is the joint binning feature set, x n is the nth feature in the feature vector set x, and || is the set concatenation symbol.

4. The joint binning label generation method for risk profile construction according to claim 1, Features: The step of performing the binning calculation on the feature set to obtain the binning value includes the following steps: Step T1, generating multiple candidate splitting thresholds for the feature set, and calculating the splitting value corresponding to each of the candidate splitting thresholds; Step T2, taking the candidate splitting threshold corresponding to the largest splitting value as the optimal splitting point; Step T3, repeatedly executing step T1 to step T2 until the number of the optimal splitting points reaches a preset value; Step T4, calculating the bin boundaries according to all the optimal splitting points; Step T5: According to the binning boundaries, the bins value and bin0 value of each of the features corresponding to each of the targets in the feature set are calculated as the binning value.

5. The method for generating joint bin labels for risk profile construction according to claim 4 is characterized in that: in, The calculation expression of the split value is: H(D j )=-Σ c∈C p(c)log2p(c), In the formula, IG(D j ,θ) is the feature set D j The corresponding split value of the candidate split threshold θ, |D| is the number of samples in D, is the feature set D j The value of feature i corresponding to target k, v k is the target k, y k is the risk judgment label corresponding to target k, C is the feature set D j The total number of features in the feature set D, p(c) is the number of features c in the feature set D j The probability of The calculation expression of the bin boundary is: In the formula for The set of bin boundary values, To sort all optimal splitting points in ascending order, The calculation expressions of the bins value and the bin0 value are: In the formula is the feature set D j The value of feature i corresponding to target k in , for The left boundary value of the u-th binning interval.

6. The joint binning label generation method for risk profile construction according to claim 1 is characterized in that: in, In step S2, the corresponding importance value is calculated according to the information gain of the feature on the node of the tree of the LightGBM model. The calculation expression of the importance value is: In the formula is the importance value of feature i in sub-feature set j, p is the total number of features in sub-feature set j, M is the total number of trees in the LightGBM model, N m is the set of nodes of the mth tree, is the indicator function, ΔL m,n is the information gain of feature i at node n of the mth tree, When node n is split using feature i, When node n does not use feature i for splitting, 7. The method for generating joint bin labels for risk profile construction according to claim 1, characterized in that: in, In step S3, the calculation expression of the importance representative feature set is: Where X lgbm_represent is the representative feature set of importance, m is the total number of sub-feature sets, X j,lgbm_top To filter the feature set corresponding to the importance of the j-th sub-feature set, All the features of the sub-feature set j are sorted from large to small according to the importance values, and a preset number of the features are selected from the beginning of the sequence to construct the importance screening feature set.

8. The method for generating joint bin labels for risk profile construction according to claim 1, characterized in that: in, In step S6, the values ​​of the features corresponding to all the targets are divided into multiple groups according to the bin0 value in the binned value of the features corresponding to the targets. The values ​​of the features corresponding to all the targets in the group have the same bin0 value, The risk judgment label is a binary classification label. The calculation expression of the woe value is: In the formula is the woe value of feature i corresponding to target k in the joint bin feature set, N bad N is the total number of risk judgment labels with bad labels corresponding to all values ​​in the group where the value of feature i corresponding to target k in the joint binning feature set belongs, good For target k in sub-feature set D j The total number of risk judgment labels whose labels are good categories in the group where the corresponding feature i value belongs, N total is the total number of values ​​of feature j corresponding to target k in the joint bin feature set.

9. A joint binning label generation device for risk profile construction, used to generate label results related to risk profiles corresponding to each target based on risk-related data, wherein the risk-related data includes values ​​of multiple different features corresponding to each target and risk judgment labels corresponding to each target, characterized in that: include: A subset division module divides all the features into multiple categories and generates a sub-feature set corresponding to each category, wherein the sub-feature set includes at least one feature and all the values ​​corresponding to the feature; An importance value calculation module, for each of the sub-feature sets, calculates the importance value of each of the features according to the number of the features in the sub-feature set and the values ​​corresponding to the features; An importance representative feature set generation module selects a plurality of features from all the sub-feature sets according to the importance values ​​to construct an importance representative feature set; A binning feature set generation module performs binning calculation on the importance representative feature set to obtain a binning value of each feature corresponding to each target in the importance representative feature set, and constructs a binning feature set based on all the binning values; A joint binning feature set generation module combines each of the features in the binning feature set to obtain a joint binning feature set; The woe calculation module performs the binning calculation on the joint binning feature set to obtain the binning values ​​of each feature corresponding to each target in the joint binning feature set, and calculates the woe values ​​of each feature corresponding to each target in the joint binning feature set based on all the binning values ​​corresponding to the joint binning feature set and the risk judgment label. Among them, the label result of the target is each of the features of the target in the joint bin feature set and its corresponding woe value.

Images